Using wavelet transform of hyperspectral reflectance curves for automated monitoring of Imperata cylindrica (cogongrass)

A wavelet-based automated classification system for detecting cogongrass from other weeds is presented. The ability of the system to detect the differences in hyperspectral reflectance curves of cogongrass and five other weeds is investigated, and the temporal effect of season is also studied. The system´s detection/classification accuracy is used to evaluate the system performance. The experimental results show that with the use of the wavelet transform for feature extraction, the classification performance is very promising, particularly when compared to more traditional data reduction methods such as the principal component transform. The experimental results also show that late summer to early fall is the best time for using the wavelet-based system to detect cogongrass from other weeds